Quantum structure and spacetime
Vincent Lam
Department of Philosophy, University of Lausanne
Abstract
The aim of this paper is twofold. In the first part, it clarifies the nature of the wave function
within the framework of the primitive ontology approach to quantum mechanics using the tools
of ontic structural realism. In the second part, it critically discusses the primitive ontological
move of postulating from the start matter localized in spacetime as the ultimate referent for
quantum theory, in particular in the case where this latter is applied to the general relativistic
gravitational field.
1
Introduction
There is a recently much discussed approach to the ontology of quantum mechanics (QM) according
to which the theory—and possibly any fundamental physical theory—is ultimately about entities in
3-dimensional space (or 4-dimensional spacetime) and their temporal evolution. Such an ontology
postulating from the start matter localized in ‘usual’ physical space or spacetime (by contrast to
e.g. a wave function on a high dimensional configuration space as within the framework of ‘wave
function realism’) is called ‘primitive ontology’ in the recent literature on the topic. According
to the proponents of the primitive ontology approach to QM, it is the best (realist) way to avoid
the main difficulty of wave function realism: there is no ‘illusion’ or ‘appearance’ of matter in 3dimensional space to be explained, since this fact is simply postulated from the start as the referent
of the theory (i.e. as what the theory is fundamentally about). Of course, this ‘postulate’ is not
the whole story: the theory in question has to specify how matter is instantiated in 3-dimensional
space (particles, fields, strings, loops, . . . ) and how it evolves in time. And that’s where things
get interesting: in all the proposed primitive ontologies for QM (the paradigmatic examples are of
course BM and versions of the theory of Ghirardi, Rimini and Weber (GRW)), the wave function
plays a central and crucial role in the time evolution of the primitive ontology and in the account
of quantum non-locality. An important and difficult task for the primitive ontologist is therefore
to elucidate what the nature of the wave function can be in this context.
The aim of this contribution is twofold. First, it aims to clarify the nature of the wave function
within the framework of the primitive ontology approach to QM using the tools of a recently
much debated conception in the metaphysics of contemporary fundamental physics, namely ontic
structural realism (OSR). Second, this article aims at discussing the very ontological move of
postulating from the start matter localized in spacetime as the ultimate referent for any fundamental
physical theory. In particular, this move seems debatable within the framework of the development
of a quantum theory for the general relativistic gravitational field, where the nature of spacetime
and its relationship to matter constitute central open issues. We will also more generally discuss
the role of spacetime in the primitive ontology approach and in the OSR conception.
1
2
The main motivation for a primitive ontology
A primitive ontology for QM specifies explicitly what the theory is fundamentally about, i.e. what
there is in the world according to QM, in terms of material entities localized in 3-dimensional
space (or 4-dimensional spacetime) and their dynamics. There are different such specifications, in
particular within two of the three standard realist conceptions of QM that take the measurement
problem seriously: particles following continuous deterministic trajectories within the framework
of BM, a continuous stochastic mass density field or stochastic point-like events (‘flashes’) within
GRW, giving rise to GRWm and GRWf respectively (see Allori et al. 2008 and references therein).
A primitive ontology in the context of the Everett (or ‘many-worlds’) framework can possibly be
defined (for instance in terms of a deterministic mass density field, see Allori et al. 2011) but its
meaning is less transparent.
The main motivation for specifying a primitive ontology for QM is rather straightforward: it
provides a powerful and generic explanatory framework within which familiar macroscopic objects
localized in 3-dimensional space and their (classical) behavior can be understood in terms of the
behavior of (possibly fundamental) microscopic entities that are also localized in 3-dimensional
space (in particular, there is an explicit connection between the behavior of these microscopic
entities and what can be observed at the macroscopic level, for instance in terms of measurement
outcomes). Obviously, the details of this account depend on the specific primitive ontology under
consideration; the point is that such an account in terms of a primitive ontology does not have to
bridge substantial explanatory gaps, for instance between macroscopic objects that are (or seem to
be) localized in 3-dimensional space and fundamental microscopic entities that are not (as it seems
to be the case within the framework of wave function realism; for recent discussions on primitive
ontology and its contrast to a wave function ontology, see the contributions in Ney and Albert
2013).
The primitive ontology approach to QM finds part of its roots in Bell’s notion of ‘local beables’,
which was introduced in the context of his reflections on non-locality and the measurement problem
(see the papers collected in Bell 1987). To some extent, a primitive ontology is made up of local
beables—that is, of entities that “can be assigned to some bounded spacetime region” (Bell 1987,
53)—which can be directly related to the behavior of familiar material objects, such as a measurement apparatus for instance, so that the measurement problem simply does not arise within
the framework of a primitive ontology for QM (in this sense, the main motivation for a primitive
ontology for QM is that there is simply no quantum measurement problem in this context).
3
The central role of the wave function
However, the local beables alone (in 3-dimensional physical space) have little explanatory power;
the material entities that are localized in 3-dimensional physical space and that constitute the
primitive ontology have little explanatory power alone. The explanatory power of the primitive
ontology approach to QM stems from the local beables together with their temporal development
or dynamics, which crucially relies on the wave function. As a consequence, in this context, it seems
unavoidable to accept the wave function on top of or as ‘part of’ the primitive ontology, in some
sense to be clarified (note that in principle it does seem possible to consider the primitive ontology
move within a purely Humean framework where all quantum features, including the wave function
and the features related to quantum non-locality, supervene on the entire spacetime distribution of
local beables, see Miller 2013—however, the standard difficulties related to the explanatory power
of Humeanism seem especially salient in the quantum case).
2
Let us consider BM as an illustration. Indeed, BM embodies the paradigmatic example of a
primitive ontology for QM and as such will serve us as a very convenient case study throughout this
contribution. The Bohmian particles constitute the primitive ontology (they obviously are local
beables since they always have a definite position in 3-dimensional physical space), but the temporal
evolution of the total configuration of the Bohmian particles crucially relies on the universal wave
function through the Bohmian guiding equation or equation of motion. According to this equation,
Bohmian particles continuously evolve along determinate trajectories, but such that the velocity of
each particle depends on the positions of all the other particles: strictly speaking the velocity of
each particle is a functional of the universal wave function defined on the whole configuration. In
particular, the role of the wave function in this huge dynamical interdependence is central to the
Bohmian account of quantum non-locality, that is, to its explanatory power (more on that below).
Clearly, this crucial role of the wave function is shared by the other main primitive ontologies
of QM (e.g. GRWm and GRWf); the wave function is an irreducible part of what Allori et al.
(2008) have identified as the ‘common structure’ of all the conceptions within the primitive ontology
approach to QM. The ‘common structure’ between GRWm, f and BM is that the considered theory is
fundamentally about matter in spacetime, in contrast to, e.g., a wave function in a high-dimensional
configuration space; however the point here, which is rather clear among the proponents of a
primitive ontology for QM, is that the wave function cannot be entirely dropped from the ontological
picture (see however the investigations in Dowker and Herbauts 2005). The theoretical, explanatory,
ontological importance of the wave function therefore creates a tension within the primitive ontology
approach to QM: it raises the (old) issue of the ontological status and the metaphysical nature of the
wave function within the familiar ontological picture offered by the primitive ontology framework,
that of matter localized in 3-dimensional space and evolving in time. Let us now consider what are
the standard options for understanding the wave function in this context.
4
Three categories for understanding the wave function
There are mainly three realist ways to understand the wave function as part of the ontology of QM
(see Belot 2012), each appealing to a different philosophical category and two of which being rather
common within the primitive ontology approach. The first one is the most straightforward: to
consider the wave function as a physical object on its own. It is also the most problematic from the
point of view of the primitive ontology approach. Indeed, within the framework of BM, it amounts
to recognize the wave function as a physical object, possibly not living in 3-dimensional space
but in high-dimensional configuration space, in addition to the Bohmian particles in 3-dimensional
space, thereby considerably inflating the ontology. The explanatory strength and simplicity of the
primitive ontology approach would then be significantly weakened, and some of the difficulties of
wave function realism—against which the primitive ontology approach was originally designed—
would reappear (in particular, the link between the local beables in spacetime and the wave function
existing in a different space would have to be clarified). It is therefore not surprising that this option
is commonly rejected within the framework of the primitive ontology approach to QM.
The second understanding appeals to an entirely different philosophical category: it suggests to
consider the wave function as a law-like, nomological entity, that is, not as an additional substantial,
physical entity in space and time. This interpretation of the wave function is favored by some of
the most prominent current proponents of BM, who take as a heuristic argument the analogy with
the common interpretation of the Hamiltonian on phase space within the framework of classical
mechanics (see e.g. Dürr et al. 1997). So, within this understanding and the Bohmian context, the
wave function is taken as an aspect of the Bohmian law of motion (guiding equation). However, this
3
nomological interpretation of the wave function faces an important difficulty: the wave function can
be time-dependent—a non-standard feature for a law-like entity, which requires some clarifications.
A related difficulty concerns the status of the Schrödinger equation: what is the status of a law
(the Schrödinger equation) that determines the temporal evolution of law-like entity (the wave
function)? In order to deal with these difficulties, the proponents of the nomological understanding
of the wave function within BM have deployed a strategy which contains three main components.
First, the crucial distinction between the universal wave function, i.e. the wave function of the
universe (the wave function corresponding to all the Bohmian particles in the universe), and the
effective wave functions corresponding to (Bohmian) subsystems of the universe. If the latter are
epistemically crucial (they are the ones that are dealt within standard QM as well as for predictive
and operational purposes), only the former is ontologically fundamental strictly speaking (effective
wave functions only possess a ‘derivative’ status compared to the universal wave function). Second,
and most importantly for the later parts of this article, the expectation, based on the fundamental
timelessness of the Wheeler-DeWitt equation in canonical QG, that the universal wave function is
static (and possibly unique). Third, the (informed) conjecture that the time-dependent Schrödinger
evolution is not fundamental, but only effective in the sense of only arising for subsystems and their
effective (‘time-dependent’) wave functions. Even if supported by good heuristic arguments, the
second and third component of this strategy remain speculative, making the purely nomological
interpretation of the wave function somewhat less attractive (the important point to underline at
this stage is the fact that the proponents of the nomological understanding of the wave function in
the context of the Bohmian primitive ontology appeal to the QG domain to ground their conception,
see e.g. Dürr et al. 1997, §12). Moreover, it seems that the exact ontological picture resulting
from the nomological understanding of the wave function depends on one’s metaphysical stance
with respect to laws, e.g. Humean or dispositional. There is no need to enter this venerable
metaphysical debate here. Suffice it to note that if a Humean approach in this context is clearly
not incoherent (leading to what could be called ‘quantum Humeanism’), it can be argued that
it would considerably weaken the explanatory power of the conception, which is one of the main
motivations for the primitive ontology move in the first place. For instance, within this Humean
framework, the wave function and crucial quantum features such as quantum non-locality that are
encoded in the wave function would merely supervene on the whole distribution of the relevant
local beables in the entire spacetime, rather than being anchored (an therefore ‘explained’ in some
sense) in the nature or properties of these local beables postulated by the primitive ontology (see
the comment above in section 3; see also the discussion in Esfeld et al. 2013).
The third understanding of the wave function precisely aims to do that: the idea is to interpret
the wave function in terms of the properties of—more precisely, the relations among—the local
beables. This understanding is appealing in the primitive ontology context: indeed, for example,
within the framework of BM, the wave function determines through the Bohmian equation of
motion (guiding equation) the temporal development of the local beables, that is, the velocities of
the Bohmian particles. In this perspective, it is perfectly sensible to think of the wave function
as describing a fundamental property of the Bohmian particles that determines their motion (this
description possibly not being one-to-one, see Belot 2012, 78-80 and the discussion in Esfeld et al.
2013, §4). The main worry for this understanding comes from the fact that, in this context, the
wave function encodes quantum non-locality, so that the fundamental property described by the
wave function is rather peculiar. In the Bohmian case, the (universal) wave function is defined
on the whole configuration of all Bohmian particles in the universe at a given time, so that the
temporal development (the velocity) of each particle depends strictly speaking on the positions of
all the other particles at that time through the (universal) wave function. Therefore, the wave
function actually describes a kind of holistic property of the whole configuration of particles (at a
4
given time). We discuss in the next two sections how ontic structural realism can help to clarify
the nature of such a property, that is, the nature of the wave function in the primitive ontology
approach.
5
Ontic structural realism and primitive ontology
Ontic structural realism (OSR) is a recently much debated conception in the metaphysics of contemporary fundamental physics, in particular quantum theory. As a metaphysical conception and
interpretative framework for fundamental physics, its development has been mainly motivated by
various fundamental relational physical features, in particular background independence and gaugetheoretic diffeomorphism invariance in the general relativistic domain (see e.g the contributions by
Pooley, Rickles and Stachel in Rickles et al. 2006 as well as Esfeld and Lam 2008) and permutation
invariance (together with other symmetry considerations), entanglement and non-locality in the
quantum domain (see e.g. French and Ladyman 2003, Esfeld 2004, Ladyman et al. 2007, ch.3,
Kantorovich 2009, Muller 2011, Lam 2013). The broad ontological thesis of OSR that is motivated
by these relational features can be expressed in the following way: what there is in the world at the
fundamental level (or in the cases where OSR is relevant) are physical structures, in the sense of
networks of concrete physical relations among concrete physical objects (relata), whose existence
depends in some sense on relations in which they stand and on structures they are part of (see
French 2010 for a discussion of the relevant notion of ‘existential dependence’ in this context; see
also recently Wolff 2012 and McKenzie 2013).
As mentioned above, it has been argued in the literature for some time now that OSR provides
a general interpretative framework for the generic relational features of quantum entanglement
and quantum non-locality as encoded in the violations of Bell-type inequalities. Since (of course)
quantum non-locality has to be accounted for within the primitive ontology approach, it is no
wonder that OSR is relevant in this context. Indeed, on the one hand, a primitive ontology for QM
(such as BM) provides an ontological framework within which the general OSR understanding of
quantum non-locality can be specified (in particular the relata of the relevant quantum structures
can be specified). On the other hand, OSR provides the primitive ontology approach to QM with
a convincing way to interpret the wave function and its encoding of quantum non-locality in this
context (see Esfeld 2014 for a similar point of view).
6
The wave function as a physical structure
We have seen above that there is a tension about the wave function within the primitive ontology
approach to QM (section 3): if it clearly plays a central role in the explanatory scheme of the
primitive ontology approach (in particular, in the account of non-locality), it is not easy to anchor
the wave function in the 3-dimensional physical space (or 4-dimensional spacetime) in which the
relevant local beables live—as one would expect within the primitive ontology framework. OSR
precisely provides such a spacetime anchorage for the wave function in the primitive ontology
approach to QM: in this context, the wave function can be understood as a physical structure in
spacetime whose relata are the local beables of the primitive ontology under consideration (this
spacetime anchorage is crucial from the primitive ontology point of view; we will come back to this
point below).
Let us illustrate how this account works in our case study, BM. The wave function is understood
in terms of a concrete, physical structure instantiated by the Bohmian particles. This huge network
of physical correlations (as described by the wave function) constitutes the physical ground for
5
(the explicit) quantum non-locality in BM; in this fundamental quantum structure, each particle
is strictly speaking related to all the others (in a way described by the wave function), so that
its temporal development (its velocity) depends on the positions of all the other particles (hence
the BM account of the violation of Bell-type inequalities in terms of the violation of parameter
independence); note that, as mentioned above in section 4, the notion of effective wave function
captures the operationally relevant aspects of such a huge dependence.
There are two aspects that jointly make this quantum structure, which is described in the
quantum formalism by the wave function, a structure in the OSR sense. First, the quantum
relations connecting all the Bohmian particles do not supervene on any intrinsic properties of
the particles; therefore, these quantum relations and the corresponding quantum structure are
fundamental and irreducible in the sense that they cannot be merely understood in terms of (they
cannot be ‘reduced’ to) the intrinsic properties of the relata, namely the Bohmian particles. Second,
even if some intrinsic individuality and identity can possibly be ascribed to them (e.g. in virtue
of their spacetime location, if one accepts that it can be taken as an intrinsic feature), there is a
sense in which Bohmian particles dynamically depend on the structure they are part of, through
the dependence on the positions of all the other particles. Unlike the case of Newtonian gravity
(where some structuralist dependence among all Newtonian particles also obtains), this dependence
is strictly speaking not affected by spatial distance. So, in a sense, the very existence of Bohmian
particles dynamically depend on the structure they are part of.
Furthermore, one could characterize Bohmian particles in terms some dynamical (diachronic)
identity that depends on the whole configuration of particles, that is, in terms of some non-intrinsic
(structural, contextual) identity (about the notion of non-intrinsic identity in the context of OSR,
see Lam 2014). The tension between this structural identity and the above mentioned intrinsic
identity based on the spacetime location is only apparent: besides the fact that its ‘intrinsicness’
can be put into question (it ultimately relies on the spacetime structure), this latter identity is
dynamically inert, whereas the former plays a crucial dynamical and explanatory role, in particular
in the account of quantum non-locality. Indeed, in this context, quantum non-locality is accounted
for in terms of the dynamics of the (relevant part of the) quantum structure, within which the relata
(i.e. the Bohmian particles) are interdependent; this dynamical interdependence is here precisely
encoded in the notion of dynamical (or structural, contextual) identity. In this perspective, it
seems to make sense to claim that for each Bohmian particle, the fact of being this very particle,
which includes its own trajectory and dynamical features, depends on the structure it is part of.
Moreover, this structuralist understanding receives further support from the permutation invariant
formulation of BM, called ‘identity-based’ BM (see Goldstein et al. 2005b,a), which highlights the
fact that Bohmian particles can be genuinely understood as lacking intrinsic properties altogether.
This quantum structure instantiated by the Bohmian particles possesses a fundamental and
inherently modal nature, as described by the wave function on configuration space. In particular,
this modal nature grounds all the possible temporal developments of the interdependent Bohmian
particles through the guiding equation. The modal nature and the dynamical role of the quantum
structure is crucial to the structuralist understanding of the wave function: this latter should not be
understood in terms of a bunch of dynamically and modally inert relations holding among Bohmian
particles (or in terms of a plethora of non-actual relations), but rather in terms of a concrete,
physical structure instantiated by the actual particle configuration that relates all the particles in
determining their temporal development. The universal wave function and the guiding equation
fully describe what sort of physical structure this quantum structure is and how it constraints
as a whole the temporal development of each particle—in particular, correlating the temporal
6
development of each particle with all the other particles.1
So, from the metaphysical point of view, the structuralist account of the wave function proposed
here provides a clear metaphysical basis for the holistic aspects mentioned within the framework of
the property understanding at the end of section 4; in this sense, OSR helps to clarify the nature and
the status of the wave function—more precisely: what is represented by the wave function—within
the primitive ontology approach to QM. OSR provides a way to understand the wave function
in spatio-temporal terms, as a concrete physical structure instantiated by local material entities
(beables) in spacetime. As mentioned above, this spacetime anchorage of the wave function is
crucial to the primitive approach to QM, since it is postulated from the start that the fundamental
ontology is about matter (and its properties/relations) in spacetime. In the next sections, we look
more closely at this spacetime postulate that is at the heart of the primitive ontology approach to
quantum theory, in particular when this latter is applied to the gravitational (general relativistic)
domain.
7
Primitive ontology and spacetime
As we have discussed in section 2, the main motivation for postulating a primitive ontology for
QM, that is, for postulating an ontology of material entities localized in spacetime is the explanatory power that such an ontological background provides. In particular, it allows for a classical
explanatory scheme, that is, an explanatory scheme similar to the one at work in classical physics,
where the behavior of familiar, macroscopic objects that appear localized in 3-dimensional space
(or 4-dimensional spacetime) can be explained in terms of microscopic, fundamental constituents
that are localized in 3-dimensional space too (we put aside possible questions related to the general
reductionist framework, they are not directly relevant here). In this context, the common spacetime
arena constitutes a crucial link between the manifest and the scientific image.
This specific (‘spacetime’) link, which is central to the classical explanatory scheme, is not
available to an ontology of QM that does not postulate material entities localized in 4-dimensional
spacetime at the fundamental level, such as within wave function realism (or, more precisely,
wave function monism; the distinction does not fundamentally alter the issue here). According
to this latter conception, what there is at the fundamental level is the wave function, understood
as a real, substantial entity (‘field’) living not in 4-dimensional space-time but in (on) a highdimensional space (isomorphic to what is usually called ‘configuration space’). To the extent that
the wave function is crucial to all the three standard realist interpretations of QM (i.e. Bohm,
GRW, Everett; see section 3), these latter can all be ontologically understood along the lines of
wave function realism. It is interesting to note that the debate between a primitive ontology
approach to QM and wave function realism is a debate between different ontological frameworks
underlying the three standard realist interpretations of QM (rather than a debate between these
interpretations; see also Ney and Philips 2013, §2). If, as discussed above in section 2, it does
seem that wave function realism has to bridge an important explanatory gap that is absent from
the primitive ontology explanatory scheme, it is also important to underline that the proponents
of a wave function ontology are not without explanatory resources (see e.g. Albert 1996, 2013,
Ney 2012). However, in the case of QM, it can be rather convincingly argued that the primitive
ontology approach has a clear explanatory (and maybe ontological) advantage, be it only in terms
of simplicity, where spacetime is a crucial component (similarly, spacetime implicitly plays a crucial
role in the motivation for Bell’s local beables).
1
Thanks to an anonymous referee for pressing me on this point.
7
Now, we would like to stress that the role of spacetime in elaborating an ontology for quantum
theory applied to the gravitational field as described by the general theory of relativity (GTR) is
not as clear-cut as it is in the case of QM (including standard quantum field theory). Indeed, in the
latter case, spacetime is not part of the dynamical physical systems described by the theory, where
a given fixed, non-dynamical (possibly curved) spacetime structure is therefore postulated from the
start (the “stage” over which the physics unfolds, as Rovelli 2001 puts it), so that it seems coherent
for the corresponding ontology to do as much. In the former case, things possibly are entirely
different. Indeed, under a common understanding of GTR, the gravitational field and the spacetime
structure are aspects of the same physical dynamical entity, which is described by the theory and
its dynamical equations, the Einstein field equations (so that the spacetime “stage” becomes an
“actor”). A crucial feature in this context is that there is no fixed, non-dynamical physical entity
with respect to which physical systems and their dynamics can be considered in a privileged way;
more precisely, the (metric-)gravitational field cannot be decomposed uniquely into an inertial
(non-dynamical) part plus a gravitational (dynamical) part (this is a rough characterization of
what is called ‘background independence’, a feature many think to be essential for a fundamental
theory of QG, see e.g. Smolin 2006; more precise discussions of background independence can be
found in Giulini 2007 and Rickles 2008 for instance). To the extent that this feature of background
independence is fundamental (and again, many think it is), one can legitimately expect that it is
reflected in the ontology in some way or another, so that we can already see at this classical stage
that a tension with the spacetime postulate of primitive ontology may possibly arise.
Indeed, if one tries to develop a quantum theory of the gravitational field—that is, a quantum
theory of the spacetime structure itself—within this background independent framework, it does
not seem to be justified to postulate from the start that the ontology of such a QG theory is
fundamentally about a spacetime arena within which and with respect to which physical systems
are localized. The very nature and status of spacetime and its relationship to matter are actually
central open issues, which are the focus of ongoing QG research programs. Within the framework of
many QG candidate theories, such as the ones based on the canonical quantization of GTR, various
ontological conceptions about spacetime have been argued for, including conceptions according to
which spacetime is not fundamental in some sense, and the whole debate is still very much ongoing
(the detailed arguments need not be exposed here, for a recent overview see Huggett and Wüthrich
2013). The important point is that the very status of spacetime is under debate within these
candidate theories of QG. A primitive ontology approach just does not seem to be able to do justice
to this very important debate and the very methodology underlying this approach seems flawed in
this context (at least to the extent that it amounts to assuming from the start a certain ontological
conception of spacetime in the very debate on the ontological nature and status of spacetime; see
section 9 below for further reflections in this context on the role and status of spacetime in string
theory and Bohmian approaches to QG).
There is for sure a venerable philosophical tradition according to which ontological claims about
space and time (and their relationship to matter) can be made a priori, without relying on our best
physical descriptions that are available. Without discussing the merits and the difficulties of such an
epistemic stance, suffice is to note that it is certainly not in line with the ‘naturalized metaphysics’
framework according to which our fundamental ontology of nature should be grounded in our
experimentally most successful physical theories and within which most of the current debates on
the ontology of quantum theory (including the primitive ontology approach) take place. As a side
remark, let us further note that there definitely are important and interesting debates around this
naturalized approach to analytical metaphysics (for a recent overview, see the collections of essays
in Ross et al. 2013) and in particular the role and status of spacetime within such a metaphysics of
science, but they constitute different (sometimes related) topics of investigations than the one the
8
primitive ontology conceptions (and this paper) are originally concerned with (namely, the ontology
of quantum theory).
8
Ontic structural realism and spacetime
The role of OSR in the understanding of the quantum wave function within the primitive ontology
framework is double. First, it aims to encode the fundamental relational features of the wave
function as manifested in quantum non-locality. Second, it allows to anchor the wave function in
spacetime, as a physical structure in spacetime among the relevant local beables constituting the
primitive ontology under consideration. As discussed in the last section, this spacetime anchorage
may not be relevant anymore in the QG domain, where the very status of spacetime is an open
issue. However, at least within canonical QG, the wave function (or the relevant quantum states)
might still possess fundamental relational features grounding non-locality at the QM level, so that
OSR might still provide a relevant interpretative framework in the QG context, even if it can
be argued that the very notion of non-locality (as well as the related notions of entanglement
and non-separability in their common understanding) make only sense, strictly speaking, within a
spatio-temporal framework (see Lam and Esfeld 2013). One worry seems to be that in order to be
entitled to use the notions of non-locality and entanglement, one has to spell out a differentiation
in terms of a plurality of entities that are entangled (or non-locally correlated) with each other.
The difficulty lies in the fact that such concrete differentiations are often made in spatio-temporal
terms, which may not be available in the QG domain (similarly, the notion of non-separability in
terms of which quantum entanglement can be understood seems to presuppose spatio-temporally
separated entities).
However, spatio-temporal differentiation need not be the only way to obtain a plurality of
fundamental objects. Indeed, within the OSR framework, the numerical diversity of fundamental
relata is often considered as a primitive fact—this primitive numerical diversity is then said to
be ‘contextual’ in the sense of a plurality of objects whose existence depends on the structures
they are part of (as a consequence, and in contrast to a primitive diversity of ‘isolated’ individuals
with intrinsic identity, primitive contextual diversity does not imply haeccesitism, see Ladyman
2007). In this structuralist context, there is actually a number of convincing arguments in favor
of such a primitive understanding of numerical diversity (see Lam 2014, §5). First, it obviously
dissolves the possible worries around the controversial issue of the status of numerical diversity
within the purely relational framework of OSR (in particular, it avoids the circularity concern
that is sometimes voiced against OSR). It is important to note that the issue of the status of the
numerical plurality of the fundamental entities in our ontology has to be faced in any case—whether
or not spacetime is fundamental; in the context of a fundamental spacetime background providing
the above mentioned standard spatio-temporal differentiation, the issue is then about the numerical
diversity of spacetime points or regions themselves (indeed, motivated by GTR diffeomorphism
invariance, Pooley 2006 and Esfeld and Lam 2008 suggests to take the contextual diversity of
spacetime points as a primitive fact). Second, primitive numerical diversity enables OSR to clearly
distinguish between issues related to the numerical identity of the relata and issues related to
their (in)distinguishability; more specifically, it explicitly frees OSR from any commitment to the
controversial Principle of the Identity of Indiscernibles (PII)—while remaining compatible with it,
in particular in its weak version (see Ainsworth 2011 for a discussion of the links between OSR and
PII)—and it allows for possible contextual but completely indiscernible objects, in the sense of not
even weakly discernible objects.
So, without delving into the details of this issue, it does seem that taking the numerical diversity
9
of the fundamental OSR relata as a primitive fact constitutes a coherent and convincing move (for
a recent discussion on this issue, see Lam 2014 and references therein). The important point here
is that primitive contextual diversity shows that spatio-temporal differentiation is not necessary
for having a multiplicity of fundamental objects and for attributing certain (relational) features
to them, in particular for defining an OSR structure. In this sense, it does not seem that OSR
is committed to spacetime; and in this sense, OSR might still provide a relevant interpretative
framework for the QG domain, whatever the fundamental status of spacetime turns out to be (as
already mentioned above, there are good reasons to think that a structuralist framework might still
be relevant in the QG domain, see the essays in Rickles and French 2006) .
The possibility of non-spatio-temporal OSR structures, possibly suggested by certain QG theories, raises however an important question. The OSR conception is about physical ontology, that
is, about what there is in the concrete physical world, and as such it is concerned with concrete
physical structures instantiated in the concrete physical world, in contrast to the abstract, mathematical structures of mathematical structuralism for instance. Now, what makes the structures
OSR claims are out there in the world physical structures in contrast to abstract mathematical
structures? One standard way to answer this question precisely relies on spacetime and is at the
basis of the whole primitive ontology framework: physical structures are in spacetime, whereas
mathematical ones are not. Of course, this criterion does not cover the case of the spacetime
structure itself (as described by GTR for instance) or the possible non-spatio-temporal structures
mentioned above in the QG context. In the face of this difficulty, a possible move is to simply
reject the distinction between abstract and concrete entities or structures (Ladyman et al. 2007,
§3.6). However, providing abstract mathematical structures with the same ontological status as
(what one would commonly think as) concrete physical structures does not seem very satisfactory.
It seems indeed more promising to consider the other standard way to characterize concrete entities
in contrast to abstract ones: causal efficacy. Concrete physical structures as opposed to abstract
mathematical ones can be considered to be causally efficacious in some modal sense, where the
relevant modality is encoded in the structure through different metaphysical strategies, involving
causal powers, dispositions or some inherent, objective modality (see above section 6; see also Esfeld
2009, Ladyman et al. 2007, ch. 2-3, Berenstain and Ladyman 2012 and recently French 2014, ch.
8-10). Such modality might characterize concrete physical structures possibly without relying on a
spatio-temporal framework, and so would fit the QG context. Trying to spell out this modality in
non-spatio-temporal terms within specific QG cases is an interesting and important task for future
research.
9
Conclusion and perspectives
In the first part of this contribution, we have highlighted the difficulties that arise about the
status and nature of the wave function within the framework of the primitive ontology approach to
QM, that is, within the realist interpretations of the theory—realist solutions to the measurement
problem—according to which QM is ultimately about (material) entities localized in 3-dimensional
physical space and evolving in time. We have then suggested how a structuralist understanding of
the wave function in the sense of OSR might solve these difficulties. We have mainly taken BM—the
paradigmatic example of a primitive ontology for QM—as a convenient study case for illustrating
the interpretative relevance of this structuralist understanding. In particular, the explicit nonlocality that is at the heart of BM (and to some extent, at the heart of QM in general) is naturally
understood in terms of the relational features of the relevant quantum structure. The upshot is that
the primitive ontology approach and OSR can be nicely combined to form a powerful interpretative
10
framework for QM and its fundamental feature of quantum non-locality.
In the second part of this contribution, we have discussed the limits of this framework. In
particular, the crucial role spacetime plays in the primitive ontology context seems at odds with
certain approaches to QG, although the OSR conception itself does not rely on a spatio-temporal
background. These considerations from the QG domain may actually seem too speculative, and
the proponents of the primitive ontology approach may want to argue that we should first deliver a
coherent ontology for QM—an experimentally extremely successful theory, in contrast to QG theories for the time being—what their approach convincingly does. However, the primitive ontology
move is often expected to be universal in the sense of being valid for any fundamental physical
theory (see e.g. Allori 2013b). Moreover, when it comes to ground a possible nomological status
for the wave function, the proponents of the primitive ontology approach to QM readily appeal to
the (canonical) QG domain and in particular to the timelessness of the Wheeler-DeWitt equation
(see e.g. Goldstein and Zanghı̀ 2013; see also section 4 above).
Allori (2013a, §8) very briefly mentions the status of spacetime in the QG domain (namely,
string theory) within the framework of a discussion on primitive ontology. She identifies a potential
difficulty in the fact that, similarly to the ‘configuration space’ of wave function realism, but to a
much lesser extent, the relevant space of string theory is not a 3- or 4-dimensional space but rather
a higher dimensional space. But then she quickly argues that the compactification of the extra
dimensions within sting theory makes the primitive ontology approach and its explanatory scheme
still relevant in this context.
Without entering into the details, the various string-theoretic ‘dualities’ actually make the situation much more complicated than that—in a way that quite explicitly conflicts with the primitive
ontology approach (see recently Huggett and Wüthrich 2013, §2.4 for a brief overview). Depending
on the dualities considered, even topology and dimension do not seem to be determinate, fundamental features of the world in the string-theoretic context. Of course, this latter context requires a
cautious attitude, but the important point here, as already discussed in section 7, is that, whatever
its merits and difficulties, it constitutes a further example in the QG domain where the very status
of spacetime and its relationship to matter are open issues, so that postulating an ontology of material entities (local beables) localized in 4-dimensional spacetime does not seem like the appropriate
interpretative move in this framework (in particular, with respect to these dualities, the ‘strings’
of string theory are no local beables in any way).
The basic strength of the primitive ontology approach is the fundamental requirement underlying the whole move that for any physical theory it should be clear what the theory is fundamentally
about. Primitive ontologies understood in this broader sense have been put forward in the context
of Bohmian approaches to canonical QG (see e.g. Goldstein and Teufel 2001). But it should be
clear that primitive ontology in this broader sense is nothing but ontology tout court understood
in a serious way. Within this broader framework, the role of local beables becomes less central: as
such, a serious ontology need not be made up of local beables. But a serious ontology without local
beables at the fundamental level is not a primitive ontology in the original sense; most importantly,
the powerful (and simple) explanatory scheme of the primitive ontology involving local beables
is not available to such an ontology (this is especially salient in various Bohmian approaches to
quantum field theory and canonical QG, but this is a topic for another paper). In any case, the
primitive ontology approach cannot have it both ways: it cannot stick to its successful explanatory
scheme and at the same time provide an ontological framework for QG.
Acknowledgements I am grateful to the Swiss National Science Foundation (Ambizione grant
PZ00P1 142536/1) for financial support.
11
References
Ainsworth, P. M. (2011). Ontic structural realism and the principle of the identity of indiscernibles.
Erkenntnis, 75:67–84.
Albert, D. Z. (1996). Elementary Quantum Metaphysics. In Cushing, J., Fine, A., and Goldstein,
S., editors, Bohmian Mechanics and Quantum Theory: An Appraisal, pages 277–284. Kluwer,
Dordrecht.
Albert, D. Z. (2013). Wave Function Realism. In Ney, A. and Albert, D. Z., editors, The Wave
Function: Essays in the Metaphysics of Quantum Mechanics. Oxford University Press, New York,
52-57.
Allori, V. (2013a). Contre les ontologies de la fonction d’onde: défense des ontologies primitives. In
Le Bihan, S., editor, Précis de philosophie de la physique, pages 116–140. Vuibert, Paris, English
version at philsci-archive.pitt.edu/9343.
Allori, V. (2013b). Primitive Ontology and the Structure of Fundamental Physical Theories. In
Ney, A. and Albert, D. Z., editors, he Wave Function: Essays in the Metaphysics of Quantum
Mechanics, pages 58–75. Oxford University Press, New York.
Allori, V., Goldstein, S., Tumulka, R., and Zanghı̀, N. (2008). On the Common Structure of
Bohmian Mechanics and the Ghirardi-Rimini-Weber Theory. British Journal for the Philosophy
of Science, 59:353–389.
Allori, V., Goldstein, S., Tumulka, R., and Zanghı̀, N. (2011). Many Worlds and Schrödinger’s
First Quantum Theory. British Journal for the Philosophy of Science, 62:1–27.
Bell, J. S. (1987). Speakable and unspeakable in quantum mechanics. Cambridge University Press,
Cambridge.
Belot, G. (2012). Quantum states for primitive ontologists. European Journal for Philosophy of
Science, 2:67–83.
Berenstain, N. and Ladyman, J. (2012). Ontic Structural Realism and Modality. In Landry, E.
and Rickles, D., editors, Structural Realism, pages 149–168. Springer, Dordrecht.
Dowker, F. and Herbauts, I. (2005). The status of the wave function in dynamical collapse models.
Foundations of Physics Letters, 18:499–518.
Dürr, D., Goldstein, S., and Zanghı̀, N. (1997). Bohmian mechanics and the meaning of the wave
function. In Cohen, R., Horn, M., and Stachel, J., editors, Experimental Metaphysics, pages
25–38. Kluwer, Dordrecht.
Esfeld, M. (2004). Quantum entanglement and a metaphysics of relations. Studies in History and
Philosophy of Modern Physics, 35:601–617.
Esfeld, M. (2009). The modal nature of structures in ontic structural realism. International Studies
in the Philosophy of Science, 23:179–194.
Esfeld, M. (2014). How to account for quantum non-locality: ontic structural realism and the
primitive ontology of quantum physics. Manuscript.
12
Esfeld, M. and Lam, V. (2008). Moderate structural realism about space-time. Synthese, 160:27–46.
Esfeld, M., Lazarovici, D., Hubert, M., and Dürr, D. (2013). The Ontology of Bohmian Mechanics.
British Journal for the Philosophy of Science, 64, doi:10.1093/bjps/axt019.
French, S. (2010). The interdependence of structure, objects and dependence. Synthese, 175:89–109.
French, S. (2014). The Structure of the World. Oxford University Press, Oxford.
French, S. and Ladyman, J. (2003). Remodelling structural realism: Quantum physics and the
metaphysics of structure. Synthese, 136(1):31–56.
Giulini, D. (2007). Remarks on the Notions of General Covariance and Background Independence.
In Seiler, E. and Stamatescu, I.-O., editors, Approaches to Fundamental Physics: An Assessment
of Current Theoretical Ideas, pages 105–122. Springer, Berlin.
Goldstein, S., Taylor, J., Tumulka, R., and Zanghı̀, N. (2005a). Are all particles identical? Journal
of Physics A: Mathematical and General, 38(7):1567.
Goldstein, S., Taylor, J., Tumulka, R., and Zanghı̀, N. (2005b). Are all particles real? Studies in
History and Philosophy of Science Part B: Studies in History and Philosophy of Modern Physics,
36:103–112.
Goldstein, S. and Teufel, S. (2001). Quantum Spacetime without Observers: Ontological Clarity
and the Conceptual Foundations of Quantum Gravity. In Callender, C. and Huggett, N., editors, Physics Meets Philosophy at the Planck Scale, pages 275–289. Cambridge University Press,
Cambridge.
Goldstein, S. and Zanghı̀, N. (2013). Reality and the Role of the Wave Function in Quantum
Theory. In Ney, A. and Albert, D. Z., editors, The Wave Function: Essays in the Metaphysics
of Quantum Mechanics, pages 91–109. Oxford University Press, New York.
Huggett, N. and Wüthrich, C. (2013). Emergent spacetime and empirical (in)coherence. Studies
in History and Philosophy of Modern Physics, 44:276–285.
Kantorovich, A. (2009). Ontic Structuralism and the Symmetries of Particles Physics. Journal for
General Philosophy of Science, 40:73–84.
Ladyman, J. (2007). On the identity and diversity of objects in a structure. Proceedings of the
Aristotelian Society Supplementary Volume, 81(1):45–61.
Ladyman, J., Ross, D., Spurett, D., and Collier, J. (2007). Every Thing Must Go: Metaphysics
Naturalized. Oxford University Press, Oxford.
Lam, V. (2013). The entanglement structure of quantum field systems. International Studies in
the Philosophy of Science, 27:59–72.
Lam, V. (2014). Entities without intrinsic physical identity. Erkentnnis, doi:10.1007/s10670-0149601-5.
Lam, V. and Esfeld, M. (2013). A dilemma for the emergence of spacetime in canonical quantum
gravity. Studies in History and Philosophy of Modern Physics, 44:286–293.
13
McKenzie, K. (2013). Priority and particle physics: Ontic structural realism as a fundamentality
thesis. British Journal for the Philosophy of Science, doi:10.1093/bjps/axt017.
Miller, E. (2013). Quantum entanglement, bohmian mechanics, and humean supervenience. Australasian Journal of Philosophy, doi:10.1080/00048402.2013.832786.
Muller, F. A. (2011). Withering away, weakly. Synthese, 180:223–233.
Ney, A. (2012). The Status of our Ordinary Three Dimensions in a Quantum Universe. Noûs,
64:525–560.
Ney, A. and Albert, D. Z., editors (2013). The Wave Function: Essays in the Metaphysics of
Quantum Mechanics. Oxford University Press, New York.
Ney, A. and Philips, K. (2013). Does an Adequate Physical Theory Demand a Primitive Ontology?
Philosophy of Science, 80:454–474.
Pooley, O. (2006). Points, Particles and Structural Realism. In D. Rickles, S. French and Saatsi,
J., editors, The Structural Foundations of Quantum Gravity, pages 83–120. Oxford University
Press, Oxford.
Rickles, D. (2008). Who’s Afraid of Background Independence? In Dieks, D., editor, Ontology
of Spacetime. Philosophy and Foundations of Physics Series. Vol.2, pages 133–152. Elsevier,
Amsterdam.
Rickles, D. and French, S. (2006). Quantum Gravity Meets Structuralism: Interweaving Relations
in the Foundations of Physics. In D. Rickles, S. French and Saatsi, J., editors, The Structural
Foundations of Quantum Gravity, pages 1–39. Oxford University Press, Oxford.
Rickles, D., French, S., and Saatsi, J. (2006). The Structural Foundations of Quantum Gravity.
Oxford University Press, Oxford.
Ross, D., Ladyman, J., and Kincaid, H., editors (2013). Scientific metaphysics. Oxford University
Press, Oxford.
Rovelli, C. (2001). Quantum spacetime: What do we know? In Callender, C. and Huggett, N.,
editors, Physics Meets Philosophy at the Planck Scale, pages 101–124. Cambridge University
Press, Cambridge.
Smolin, L. (2006). The Case for Background Independence. In D. Rickles, S. French and Saatsi,
J., editors, The Structural Foundations of Quantum Gravity, pages 196–239. Oxford University
Press, Oxford.
Wolff, J. (2012). Do Objects Depend on Structures? British Journal for the Philosophy of Science,
63:607–625.
14